9-bit cyclic Beckett-Gray codes (Chris Dzoba, 2026). Write-up: https://dzoba.com/blog/beckett-gray-9 FORMAT A code is written as its transition sequence: 512 digits, digit k is the bit (0..8) flipped at step k, starting from the all-zero word. Bit 0 is the least significant bit. After 512 steps the walk is back at 0. FILES FOUND-9.txt Both codes, one per line as " ", with '#' comment lines giving how and when each was found. Code 1 is a digit string; code 2 is written with spaces between the digits. code1.txt Code 1 alone: exactly the 512 digits, no newline. Its SHA-256 is the one committed in https://gist.github.com/dzoba/59763c339624bcb3b5277be767fe3493 on 2026-10-31: b3d4a2ab934852099823d7dcc4adc629549d2dd20a92a2ed2277c31607588fb9 check_queue.py Checker 1: replays the FIFO queue directly (Python 3, no dependencies). check_runs.c Checker 2: separate logic in C. It treats each stretch where a bit stays on as an interval and checks that the k-th interval to start is the k-th to end. bgevo74.cpp The search program that found both codes (evolved program #74), exactly as it was run. lean-bg9/ Lean 4 project. BG.code9_isBeckettGray proves that code 1 is a 9-bit Beckett-Gray code by decide +kernel (axioms: propext only). The Lean proof covers code 1 only. COMMANDS shasum -a 256 code1.txt # or: sha256sum code1.txt python3 check_queue.py FOUND-9.txt cc -O2 -o check_runs check_runs.c && ./check_runs FOUND-9.txt python3 check_queue.py --n 9 code1.txt clang++ -O3 -march=native -std=c++17 -o bgevo74 bgevo74.cpp ./bgevo74 --n 9 --seed 720015 --time 1200 # --time is CPU seconds; prints BEST/CERT lines, CODE on success. # On 2026-11-02 this reprinted code 2 after about 337 CPU-s; # how much search fits in a CPU-second depends on machine load. cd lean-bg9 && lake build # Lean 4 v4.32.1 (see lean-toolchain), about 100 s